def get_initial_allele_counts(self, fragment):
'''Get allele counts from the initial time point'''
import os
from hivwholeseq.patients.samples import SamplePat
for i in xrange(len(self.samples)):
sample = SamplePat(self.samples.iloc[i])
if os.path.isfile(sample.get_allele_counts_filename(fragment)):
return sample.get_allele_counts(fragment)
def get_initial_allele_counts(self, fragment):
'''Get allele counts from the initial time point'''
import os
from hivwholeseq.patients.samples import SamplePat
for i in xrange(len(self.samples)):
sample = SamplePat(self.samples.iloc[i])
if os.path.isfile(sample.get_allele_counts_filename(fragment)):
return sample.get_allele_counts(fragment)
print protein, samplename
sample = SamplePat(sample)
# NOTE: How do we find what fragment covers the protein? Well, a
# protein can happily cross fragments. Since each
# codon is independent, we should iterate over codons. We do not
# do that for efficiency reasons. Instead, we identify all potential
# fragments and split the protein into full codon chunks covered by
# a single fragment.
fragment_rois = sample.get_fragments_covered(
protein, include_coordinates=True)
refseq = sample.get_reference(protein)
fn_out = sample.get_allele_counts_filename(protein,
PCR=PCR,
qual_min=qual_min,
type='aa')
from hivwholeseq.utils.sequence import alphaa
count = np.zeros((len(alphaa), len(refseq) // 3), int)
for frroi in fragment_rois:
fragment = frroi['name']
start_fr, end_fr = frroi['fragment']
start, end = frroi['roi']
# Check that we align with codons
rf = start % 3
if rf:
start_fr += 3 - rf
start += 3 - rf
rf = end % 3
sample = SamplePat(sample)
pname = sample.patient
conss_genomewide = SeqIO.read(
get_initial_reference_filename(pname, 'genomewide'), 'fasta')
# Collect the allele counts (where possible)
acs = []
for fragment in ['F' + str(i) for i in xrange(1, 7)]:
try:
ref = ''.join(
SeqIO.read(get_initial_reference_filename(pname, fragment),
'fasta'))
ac = sample.get_allele_counts(fragment, merge_read_types=False)
acs.append((fragment, ref, ac))
except IOError:
continue
if not len(acs):
if VERBOSE >= 1:
print 'No data found: skipping'
continue
# Merge allele counts
ac = merge_allele_counts(conss_genomewide, acs, VERBOSE=VERBOSE)
if save_to_file:
fn_out = sample.get_allele_counts_filename('genomewide')
np.save(fn_out, ac)
if VERBOSE >= 1:
print 'Genomewide allele counts saved to:', fn_out
for fragment in fragments:
counts = []
for samplename, sample in samples.iterrows():
if submit:
fork_self(samplename, fragment, VERBOSE=VERBOSE, qual_min=qual_min)
continue
if VERBOSE >= 1:
print fragment, samplename
sample = SamplePat(sample)
pname = sample.patient
refseq = SeqIO.read(get_initial_reference_filename(pname, fragment), 'fasta')
fn = sample.get_mapped_filtered_filename(fragment, PCR=PCR)
if not os.path.isfile(fn):
warn('No BAM file found', NoDataWarning)
continue
count, _ = gac(fn, len(refseq), qual_min=qual_min, VERBOSE=VERBOSE)
counts.append(count)
if save_to_file:
fn_out = sample.get_allele_counts_filename(fragment, PCR=PCR,
qual_min=qual_min)
count.dump(fn_out)
if VERBOSE >= 2:
print 'Allele counts saved:', samplename, fragment
fork_self(samplename,
fragment,
VERBOSE=VERBOSE,
qual_min=qual_min)
continue
if VERBOSE >= 1:
print fragment, samplename
sample = SamplePat(sample)
pname = sample.patient
refseq = SeqIO.read(
get_initial_reference_filename(pname, fragment), 'fasta')
fn = sample.get_mapped_filtered_filename(fragment, PCR=PCR)
if not os.path.isfile(fn):
warn('No BAM file found', NoDataWarning)
continue
count, _ = gac(fn, len(refseq), qual_min=qual_min, VERBOSE=VERBOSE)
counts.append(count)
if save_to_file:
fn_out = sample.get_allele_counts_filename(fragment,
PCR=PCR,
qual_min=qual_min)
count.dump(fn_out)
if VERBOSE >= 2:
print 'Allele counts saved:', samplename, fragment
print protein, samplename
sample = SamplePat(sample)
# NOTE: How do we find what fragment covers the protein? Well, a
# protein can happily cross fragments. Since each
# codon is independent, we should iterate over codons. We do not
# do that for efficiency reasons. Instead, we identify all potential
# fragments and split the protein into full codon chunks covered by
# a single fragment.
fragment_rois = sample.get_fragments_covered(protein,
include_coordinates=True)
refseq = sample.get_reference(protein)
fn_out = sample.get_allele_counts_filename(protein, PCR=PCR,
qual_min=qual_min,
type='aa')
from hivwholeseq.utils.sequence import alphaa
count = np.zeros((len(alphaa), len(refseq) // 3), int)
for frroi in fragment_rois:
fragment = frroi['name']
start_fr, end_fr = frroi['fragment']
start, end = frroi['roi']
# Check that we align with codons
rf = start % 3
if rf:
start_fr += 3 - rf
start += 3 - rf
rf = end % 3
for samplename, sample in samples.iterrows():
if VERBOSE >= 1:
print samplename
sample = SamplePat(sample)
pname = sample.patient
conss_genomewide = SeqIO.read(get_initial_reference_filename(pname, 'genomewide'), 'fasta')
# Collect the allele counts (where possible)
acs = []
for fragment in ['F'+str(i) for i in xrange(1, 7)]:
try:
ref = ''.join(SeqIO.read(get_initial_reference_filename(pname, fragment), 'fasta'))
ac = sample.get_allele_counts(fragment, merge_read_types=False)
acs.append((fragment, ref, ac))
except IOError:
continue
if not len(acs):
if VERBOSE >= 1:
print 'No data found: skipping'
continue
# Merge allele counts
ac = merge_allele_counts(conss_genomewide, acs, VERBOSE=VERBOSE)
if save_to_file:
fn_out = sample.get_allele_counts_filename('genomewide')
np.save(fn_out, ac)
if VERBOSE >= 1:
print 'Genomewide allele counts saved to:', fn_out